From 8fceb2fb50244544f4ed5ebc80a3919f25e2f685 Mon Sep 17 00:00:00 2001 From: Kevin Mingtarja Date: Thu, 10 Oct 2024 00:44:22 +0700 Subject: [PATCH 1/3] first draft for hosted models --- hosted-models.mdx | 43 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 43 insertions(+) diff --git a/hosted-models.mdx b/hosted-models.mdx index c8db996e..20e12a05 100644 --- a/hosted-models.mdx +++ b/hosted-models.mdx @@ -2,3 +2,46 @@ title: Hosted Models description: "" --- + +With Hypermode, you can simply pick a model from [Hugging Face](https://huggingface.co/), and we will manage and run it for you. + +Here (TODO) are the list of Hugging Face models we support today + +## Setup + +To use a Hypermode-hosted model, set `host: "hypermode"`, `provider: "hugging-face"`, and set `sourceModel` to be +the model name as specified on Hugging Face. + +```json hypermode.json +{ + ... + "models": { + "text-generator": { + "sourceModel": "meta-llama/Llama-3.1-8B-Instruct", + "provider": "hugging-face", + "host": "hypermode" + } + } + ... +} +``` + +## Mode of deployment + +Internally, we run our most popular models as multi-tenant, shared among different users. + +By default, if the model you use is available as a shared model, your inferences will run against these shared models. +You can override this default behavior by setting `dedicated: true` on your model in the manifest. + +Otherwise, if the model you use isn't available as shared, we will spin up a dedicated instance of the model for you. + + + These are the models available today as shared: - + [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) + - + [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2) + - + [AntoineMC/distilbart-mnli-github-issues](https://huggingface.co/AntoineMC/distilbart-mnli-github-issues) + - + [distilbert/distilbert-base-uncased-finetuned-sst-2-english](https://huggingface.co/distilbert/distilbert-base-uncased-finetuned-sst-2-english) + From 5ac65cd2c1a1f9bc1f57bcfad61eea1c975f4ae2 Mon Sep 17 00:00:00 2001 From: Kevin Mingtarja Date: Thu, 10 Oct 2024 00:45:34 +0700 Subject: [PATCH 2/3] fix list From 313f4eb3db9cb8e925c76c0f4dc5feebcfe8f3b6 Mon Sep 17 00:00:00 2001 From: Kevin Mingtarja Date: Thu, 10 Oct 2024 00:46:40 +0700 Subject: [PATCH 3/3] fix list --- hosted-models.mdx | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-) diff --git a/hosted-models.mdx b/hosted-models.mdx index 20e12a05..edca83a6 100644 --- a/hosted-models.mdx +++ b/hosted-models.mdx @@ -36,12 +36,9 @@ You can override this default behavior by setting `dedicated: true` on your mode Otherwise, if the model you use isn't available as shared, we will spin up a dedicated instance of the model for you. - These are the models available today as shared: - - [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) - - - [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2) - - - [AntoineMC/distilbart-mnli-github-issues](https://huggingface.co/AntoineMC/distilbart-mnli-github-issues) - - - [distilbert/distilbert-base-uncased-finetuned-sst-2-english](https://huggingface.co/distilbert/distilbert-base-uncased-finetuned-sst-2-english) + These are the models available today as shared: + - [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) + - [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2) + - [AntoineMC/distilbart-mnli-github-issues](https://huggingface.co/AntoineMC/distilbart-mnli-github-issues) + - [distilbert/distilbert-base-uncased-finetuned-sst-2-english](https://huggingface.co/distilbert/distilbert-base-uncased-finetuned-sst-2-english)